dan_the_3rd

> Could this in anyway lead to endless loops? I don't know how, but my spider sense is tingling. Oh I meant just download the built binary files (not the...

I'm trying to figure out internally how we can store that on s3 - it's thanksgiving in the US so hopefully we will have more news to share next week.

It looks a bit more complicated than I expected - for now I think we could upload to [pypi](https://pypi.org/project/xformers/) directly. I think a first step would be to upload files...

I've just added a secret "PYPI_TOKEN" (starting with `pypi-`) that should have access to the xformers pypi project. We can try like this and iterate. The token [should be configured...

Sorry for the lack of updates recently, with the pytorch conference it took some time to get the information I needed. So apparently there is this way of building libraries...

That is right. @fmassa how is pypi upload handled in torchvision? I assume we should be able to rely on NOVA to create the builds and then upload however we...

Oh wow that was fast :o Thanks a lot for doing this! If you happen to have any issue with NOVA, let me know and we can have them update...

> Otherwise, we can also upload manually with twine if you want to give it a shot with a test repo. It might be the way forward for the time...

Oh nice! Let's only upload for the latest pytorch stable (1.13 at the moment), with cuda 11.7 for instance.

> The second question is regarding the version, do we want to upload `dev` versions? or only tagged commits? and how should the version string look like? The optimal solution...